◆ About
About Michael Belli
I’ve spent 20+ years turning data into decisions companies can act on, building the models, and leading the analysts who build them. My work has centered on the questions where the stakes are highest: which customers are worth acquiring, which marketing spend actually pays, and whether a model you’re about to bet a budget on is telling you the truth.
I work on the client’s side. Sometimes that means building the model. Sometimes it means being the independent expert who makes sure the firm you hired built a good one. Either way, the job is the same: get the decision right.
Two ways I help
Direct-mail targeting models
For high-value acquisition — subscription services and big-ticket purchases. I build the model that tells you who to mail and where to stop, so your budget reaches the households worth reaching.
Independent oversight
For companies commissioning modeling work — an experienced modeler in the room representing you while a vendor builds your MMM, targeting model, or predictive analytics.
What I believe about models
A model isn’t good because it fits the past. It’s good because it improves the next decision. I’ve watched impressive-looking models — high accuracy, clean charts — quietly send budgets the wrong way, because they were validated against the wrong thing. I judge a model by whether it gets the direction and magnitude of a real decision right, and I say so plainly, even when the answer is “this won’t pay — don’t do it.”
That plain-spokenness is the whole value. You’re not hiring me to be impressed. You’re hiring me to be right, and to tell you when I’m not sure.
Background
My path here didn’t start in marketing. I began analyzing Medicare claims data for healthcare quality audits, then spent several years predicting customer lifetime value across a telecom company’s product lines. After a detour managing real estate investments, I came back to analytics for good — e-commerce web analytics and campaign measurement, then a stretch building the media-mix models that guided national ad spend for a major automaker.
From there I spent nearly eight years building the analytics function at a national senior living company from the ground up: hiring the team, setting the standards, and using marketing mix modeling and customer lifetime value estimates to guide a marketing budget north of $10 million. Most recently, I’ve applied machine learning to categorize products for a global beauty brand and built the direct-mail targeting models that lifted qualified lead generation 32% for a senior living company.
I hold a B.A. in Economics (High Honors) and a B.A. in History from Michigan State University, and completed my M.S. in Economics/Econometrics at the University of Wisconsin–Madison, where I was ABD toward a Ph.D. That training is why I default to asking whether a model’s result would hold up under scrutiny, not just whether it looks good in a slide.
See an end-to-end modeling project ↗ · LinkedIn ↗ · Download CV ↗
Let’s talk
Tell me what you’re working on — I’ll tell you honestly whether and how I can help.